Week Five - Repeated Measures ANOVA Flashcards
What is a repeated measures ANOVA?
Where same subjects are tested repeatedly
Advantages of RMA?
Economy of subjects
Each subject acts as their own control (reducing error variance - making test more sensitive)
Disadvantages of RMA?
ORDER EFFECTS (learning or fatigue) - can compensate by counterbalancing
Differential CARRYOVER effect (can only combat by going to between subject)
Explain the key influence of variance in RM?
We can pull out and explain a larger amount of variance in RM and therefore have a smaller error term (easier to identify sig)
What is an additional assumption of RM?
Sphericity
What is sphericity?
It requires the variance of the differences between each of the levels to be the same
When is Sphericity always met?
When there is only two levels of an IV - more levels means not met almost always
What happens when sphericity is not met?
The anova becomes too liberal/generous at identifying differences as significant (calling sig when its prob not)
What do we do if sphericity has been violated?
We need to correct for it by adjusting the degrees of freedom in the F stat by a correction factor. - done with GG
What test is there to determine whether you need to make a epsilon correction?
Mauchly (p value greater than 0.05 means sphericity has been met)
What do n2 and n2p represent?
The variation %
What is the main thing that counterbalancing does?
Distributes the effects of RM equally across conditions
What is a limitation of counterbalancing?
We are limited to sample sizes that are multiples of the possible number of orders
Why would we avoid counterbalancing for large sample sizes?
It would undermine the benefit of having RM
Benefit ot latin designs?
More economical as they counterbalance using a selection of possible orders rather than all possible orders